Accuracy of Assessing the Mandibular Canal on Cone-Beam Computed Tomography: A Validation Study

Towards clinically applicable automated mandibular canal segmentation on CBCT

OBJECTIVES

To develop a deep learning-based system for precise, robust, and fully automated segmentation of the mandibular canal on cone beam computed tomography (CBCT) images.

METHODS

The system was developed on 536 CBCT scans (training set: 376, validation set: 80, testing set: 80) from one center and validated on an external dataset of 89 CBCT scans from 3 centers. Each scan was annotated using a multi-stage annotation method and refined by oral and maxillofacial radiologists. We proposed a three-step strategy for the mandibular canal segmentation: extraction of the region of interest based on 2D U-Net, global segmentation of the mandibular canal, and segmentation refinement based on 3D U-Net.

RESULTS

The system consistently achieved accurate mandibular canal segmentation in the internal set (Dice similarity coefficient [DSC], 0.952; intersection over union [IoU], 0.912; average symmetric surface distance [ASSD], 0.046 mm; 95% Hausdorff distance [HD95], 0.325 mm) and the external set (DSC, 0.960; IoU, 0.924; ASSD, 0.040 mm; HD95, 0.288 mm).

CONCLUSIONS

These results demonstrated the potential clinical application of this AI system in facilitating clinical workflows related to mandibular canal localization.

CLINICAL SIGNIFICANCE

Accurate delineation of the mandibular canal on CBCT images is critical for implant placement, mandibular third molar extraction, and orthognathic surgery. This AI system enables accurate segmentation across different models, which could contribute to more efficient and precise dental automation systems.

Radiological assessment of the inferior alveolar canal and mental foramen using cone beam computed tomography for pre-operative evaluation of surgeries in the mandible: A single-center five-year retrospective study

Introduction

Cone beam computed tomography (CBCT) plays a significant role in studying the anatomical structures of the mandible.

Aim

This retrospective study aimed to assess the role of CBCT at the pre-assessment stage of mandibular surgery.

Materials and methods

A total of 250 CBCT images were collected. The inferior alveolar canal (IAC) and mental foramen (MF) were measured bilaterally to the mandibular teeth apexes, including molars, premolars, and canines, to the buccal, lingual cortical bone, and to the inferior border of the mandible.

Results

There were no differences in the average number of extracted teeth between the right and left sides (P-value > 0.05, median = three teeth). It was noticed that the average measures of each point of the IAC and MF on the right side were closely matched to the similar point on the left side (P-value > 0.05). T-tests showed that there were differences between males and females on the M2 and M3 on the right side (P-value < 0.05) and on the M1, M2, and M3 on the left side (P-value < 0.05). Using one-way ANOVA tests, results showed that there were some differences in measures at P0 (F = 3.376, P-value = 0.003), P4 (F = 3.782, P-value = 0.001) on the right side, and at P3 (F = 5.620, P-value = 0.019) on the left side of the mandible.

Conclusions

There were no significant differences in IAC and MF positions between the right and left sides. However, between males and females, MF measurements showed significant differences on some points on the right and left sides. Although the history of extracted teeth showed no statistically significant difference in the location of IAC and MF, the number of extracted teeth showed an effect in the IAC position on the right and left sides, but not with the MF.

Accurate mandibular canal segmentation of dental CBCT using a two-stage 3D-UNet based segmentation framework

OBJECTIVES

The objective of this study is to develop a deep learning (DL) model for fast and accurate mandibular canal (MC) segmentation on cone beam computed tomography (CBCT).

METHODS

A total of 220 CBCT scans from dentate subjects needing oral surgery were used in this study. The segmentation ground truth is annotated and reviewed by two senior dentists. All patients were randomly splitted into a training dataset (n = 132), a validation dataset (n = 44) and a test dataset (n = 44). We proposed a two-stage 3D-UNet based segmentation framework for automated MC segmentation on CBCT. The Dice Similarity Coefficient (DSC) and 95% Hausdorff Distance (95% HD) were used as the evaluation metrics for the segmentation model.

RESULTS

The two-stage 3D-UNet model successfully segmented the MC on CBCT images. In the test dataset, the mean DSC was 0.875 ± 0.045 and the mean 95% HD was 0.442 ± 0.379.

CONCLUSIONS

This automatic DL method might aid in the detection of MC and assist dental practitioners to set up treatment plans for oral surgery evolved MC.

Development and validation of a novel artificial intelligence driven tool for accurate mandibular canal segmentation on CBCT

OBJECTIVES

The objective of this study is the development and validation of a novel artificial intelligence driven tool for fast and accurate mandibular canal segmentation on cone beam computed tomography (CBCT).

METHODS

A total of 235 CBCT scans from dentate subjects needing oral surgery were used in this study, allowing for development, training and validation of a deep learning algorithm for automated mandibular canal (MC) segmentation on CBCT. Shape, diameter and direction of the MC were adjusted on all CBCT slices using a voxel-wise approach. Validation was then performed on a random set of 30 CBCTs - previously unseen by the algorithm - where voxel-level annotations allowed for assessment of all MC segmentations.

RESULTS

Primary results show successful implementation of the AI algorithm for segmentation of the MC with a mean IoU of 0.636 (± 0.081), a median IoU of 0.639 (± 0.081), a mean Dice Similarity Coefficient of 0.774 (± 0.062). Precision, recall and accuracy had mean values of 0.782 (± 0.121), 0.792 (± 0.108) and 0.99 (± 7.64×10^-05) respectively. The total time for automated AI segmentation was 21.26 s (±2.79), which is 107 times faster than accurate manual segmentation.

CONCLUSIONS

This study demonstrates a novel, fast and accurate AI-driven module for MC segmentation on CBCT.

CLINICAL SIGNIFICANCE

Given the importance of adequate pre-operative mandibular canal assessment, Artificial Intelligence could help relieve practitioners from the delicate and time-consuming task of manually tracing and segmenting this structure, helping prevent per- and post-operative neurovascular complications.

Determining the Accuracy of the Mandibular Canal Region in 3D Biomodels Fabricated from CBCT Scanned Data: A Cadaveric Study

OBJECTIVE

To validate the accuracy of the mandibular canal region in 3D biomodel produced by using data obtained from Cone-Beam Computed Tomography (CBCT) of cadaveric mandibles.

METHODS

Six hemi-mandible samples were scanned using the i-CAT CBCT system. The scanned data was transferred to the OsiriX software for measurement protocol and subsequently into Mimics software to fabricate customized cutting jigs and 3D biomodels based on rapid prototyping technology. The hemi-mandibles were segmented into 5 dentoalveolar blocks using the customized jigs. Digital calliper was used to measure six distances surrounding the mandibular canal on each section. The same distances were measured on the corresponding cross-sectional OsiriX images and the 3D biomodels of each dentoalveolar block.

RESULTS

Statistically no significant difference was found when measurements from OsiriX images and 3D biomodels were compared to the "gold standard" -direct digital calliper measurement of the cadaveric dentoalveolar blocks. Moreover, the mean value difference of the various measurements between the different study components was also minimal.

CONCLUSION

Various distances surrounding the mandibular canal from 3D biomodels produced from the CBCT scanned data was similar to that of direct digital calliper measurements of the cadaveric specimens.

Accuracy of linear measurements on CBCT images related to presurgical implant treatment planning: A systematic review

OBJECTIVE

The aim of this systematic review was to identify, review, analyze, and summarize available evidence on the accuracy of linear measurements when using maxillofacial cone beam computed tomography (CBCT) specifically in the field of implant dentistry.

MATERIAL AND METHODS

The search was undertaken in April 2017 in the National Library of Medicine database (Medline) through its online site (PubMed), followed by searches in the Cochrane, EMBASE, ScienceDirect, and ProQuest Dissertation and Thesis databases. The main inclusion criterion for studies was that linear CBCT measurements were performed for quantitative assessment (e.g., height, width) of the alveolar bone at edentulous sites or measuring distances from anatomical structures related to implant dentistry. The studies should compare these values to clinical data (humans) or ex vivo and/or experimental (animal) findings from a "gold standard."

RESULTS

The initial search yielded 2,516 titles. In total, 22 studies were included in the final analysis. Of those, two were clinical and 20 ex vivo investigations. The major findings of the review indicate that CBCT provides cross-sectional images that demonstrate high accuracy and reliability for bony linear measurements on cross-sectional images related to implant treatment. A wide range of error has been reported when performing linear measurements on CBCT images, with both over- and underestimation of dimensions in comparison with a gold standard. A voxel size of 0.3 to 0.4 mm is adequate to provide CBCT images of acceptable diagnostic quality for implant treatment planning.

CONCLUSIONS

CBCT can be considered as an appropriate diagnostic tool for 3D preoperative planning. Nevertheless, a 2 mm safety margin to adjacent anatomic structures should be considered when using CBCT. In clinical practice, the measurement accuracy and reliability of linear measurements on CBCT images are most likely reduced through factors such as patient motion, metallic artefacts, device-specific exposure parameters, the software used, and manual vs. automated procedures.

Potential reformatting errors in cone-beam computed tomography for dentoalveolar surgery: a cautionary tale

We describe the removal of a lower second molar tooth in which preoperative cone-beam computed tomography (CBCT) showed that the inferior alveolar nerve (IAN) was encased in the distal apex of the root of the tooth. During operation the nerve was found to be entirely separate from the apex of the root and not involved. With the wider use of CBCT in the treatment planning of dentoalveolar surgery, this case represents a cautionary tale to the clinician on reliance on clinical imaging and software in guiding the decision making process.

Comparing Two Diagnostic Procedures in Planning Dental Implants to Support a Mandibular Free-Ending Removable Partial Denture

BACKGROUND

The use of a cone beam computed tomography (CBCT) for the preoperative implant planning is increasing. A clear guideline is needed in which cases of CBCT is essential.

PURPOSE

In this study, two imaging modalities (panoramic radiograph and CBCT) are compared in preoperative implant planning in the severely resorbed mandible and the influence on the observers assessments.

MATERIALS AND METHODS

Thirty-four consecutive patients with bilateral edentulous regions in the mandible were included. The feasibility of implant placement in the premolar and molar region was judged by three observers on basis of casts either with a panoramic radiograph or a CBCT.Cohen's kappa, sensitivity and specificity rates, odds of agreement and disagreement as well as the odds ratios (ORs, ratio between odds of agreement and disagreement) were calculated per observer and overall for all observers assuming the majorities agreement as the prevailing opinion.

RESULTS

Overall outcome for premolar region revealed true-positive and true-negative rates of 90% and 0%, respectively, with Cohen's kappa (κ) = -0.04. The ORs for the three observers varied between 2.6 and 158.8, with an overall OR = 76.For the molar region, overall true-positive and true-negative rates were 65% and 22% respectively, with Cohen's κ = 0.68, representing a reasonable amount of agreement. Sensitivity and specificity as well as the ORs for individual observers were fairly consistent with an overall OR = 43.

CONCLUSION

Implant placement in the resorbed posterior mandible can be well assessed with a cast in combination with a panoramic radiograph in the vast majority of the cases. Misclassification amounts to approximately 10% to 13%. In all cases of misclassification, a critical bone height, or an unclear course of the mandibular nerve or a knife edge ridge was present. In these cases, the use of a CBCT is justified.